Table 2 Sensitivity analysis of the association of conjunctival hyperaemia and the use of latanoprost, travoprost and bimatoprost
| Latanoprost vs bimatoprost | |||
| Statistical model | No of studies | No of patients | OR (95% CI) |
| Fixed effects | 8 | 1240 | 0.32 (0.24 to 0.42) |
| Random effects | 8 | 1240 | 0.32 (0.24 to 0.42) |
| Analyses excluding | |||
| Dubiner et al25 | 7 | 1197 | 0.31 (0.24 to 0.41) |
| Gandolfi et al24 | 7 | 1008 | 0.33 (0.24 to 0.44) |
| Noecker et al26 | 7 | 971 | 0.32 (0.23 to 0.44) |
| Parrish et al32 | 7 | 968 | 0.29 (0.21 to 0.40) |
| Walters et al27 | 7 | 1164 | 0.32 (0.24 to 0.43) |
| Konstas et al28 | 7 | 1198 | 0.33 (0.25 to 0.43) |
| Dirks et al29 | 7 | 1180 | 0.32 (0.24 to 0.42) |
| Konstas et al30 | 7 | 994 | 0.34 (0.25 to 0.45) |
| Latanoprost vs travoprost | |||
| Statistical model | No of studies | No of patients | OR (95% CI) |
| Fixed effects | 6 | 1118 | 0.51 (0.39 to 0.67) |
| Random effects | 6 | 1118 | 0.51 (0.39 to 0.67) |
| Analyses excluding | |||
| Netland et al31 | 5 | 523 | 0.54 (0.36 to 0.80) |
| Parrish et al32 | 5 | 844 | 0.46 (0.33 to 0.64) |
| Parmaksiz et al33 | 5 | 1084 | 0.53 (0.40 to 0.69) |
| Chen et al34 | 5 | 1045 | 0.51 (0.39 to 0.67) |
| Garcia-Feijoo et al35 | 5 | 1056 | 0.51 (0.39 to 0.68) |
| Konstas et al36 | 5 | 1038 | 0.53 (0.40 to 0.71) |
